metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: response_a
dtype: string
- name: response_b
dtype: string
- name: more_reasonable
dtype: string
splits:
- name: train
num_bytes: 1392007
num_examples: 3980
download_size: 205770
dataset_size: 1392007
configs:
- config_name: default
data_files:
- split: train
path: data/train-*